DRR generation and enhancement using a dedicated graphics device

• US 20060291710A1
• Filed: 06/23/2005
• Published: 12/28/2006
• Est. Priority Date: 06/23/2005
• Status: Active Grant

• Current Assignee: Accuray Incorporated
• Original Assignee: Accuray Incorporated
• Inventors: Fu, Dongshan, Wang, Bai
• Granted Patent: US 7,330,578 B2
• Time in Patent Office: Days
• US Class Current: 382/131
• CPC Class Codes: G06T 15/08 Volume rendering

Abstract

DRR generation and enhancement using a dedicated graphics device are described herein. In one embodiment, an example of a process for generating DRR (digitally reconstructed radiography) images includes, but is not limited to, loading a volume rendering program into a graphics device, the volume rendering program representing a predetermined algorithm according to a non-linear attenuation model. In response to 3D scan data, the graphics device is invoked to execute the volume rendering program to perform at least a portion of volume rendering operations on at least a portion of the 3D scan data, which may include modifying the 3D scan data according to the predetermined algorithm to compensate for a difference between a first attenuation of an object with respect to a second attenuation associating a known intensity of the object. Other methods and apparatuses are also described.

Claims

1. A computer implemented method for generating DRR (digitally reconstructed radiography) images, the method comprising:

loading a volume rendering program into a graphics device, the volume rendering program representing a predetermined algorithm according to a non-linear attenuation model; and

in response to 3D scan data, invoking the graphics device to executing the volume rendering program to perform at least a portion of volume rendering operations on at least a portion of the 3D scan-data, including modifying the 3D scan data according to the predetermined algorithm to compensate for a difference between a first attenuation of an object with respect to a second attenuation associating a known intensity of the object.

2. The method of claim 1, wherein the volume rendering operations further comprise:

casting a plurality of hypothetical rays through the modified 3D scan data from at least one of the known intensity, a known orientation, and an angle;

integrating the 3D scan data along with each hypothetical ray; and

projecting integrated values of 3D scan data onto an image plane.

3. The method of claim 1, wherein the 3D scan data comprises at least one of CT, MRI, PET, and ultrasound scan data.

4. The method of claim 1, wherein the 3D scan data comprises a plurality of CT numbers representing an image intensity of corresponding 3D CT voxels, wherein each CT voxel represents a corresponding volume element of the object, and wherein each CT number represents an attenuated intensity of an X-ray CT beam that has been generated at a CT scan energy level and that has traversed the corresponding volume element of an anatomical region.

5. The method of claim 4, wherein each CT voxel is disposed within one of a plurality of axial voxel slices, each axial voxel slice representing a corresponding axial slice of the object.

6. The method of claim 5, wherein integrating the CT numbers comprises:

performing a tri-linear interpolation for the CT voxels encountered by the ray; and

for each voxel of interest, performing a one-dimensional polynomial interpolation over the voxel of interest and for voxels on each adjacent voxel slice.

7. The method of claim 1, wherein the predetermined algorithm is performed according to a following algorithm:

C(x,y,z)=aC₀(x,y,z)e^{bC0(x y z)}

wherein C(x,y,z) represents the modified CT number of a 3D CT voxel having a location (x,y,z), wherein a and b represent weighting coefficients, and wherein C₀(x,y,z) represents the un-modified CT number based on a linear attenuation model of a 3D CT voxel having a location (x,y,z).

8. The method of claim 1, wherein the at least a portion of the volume rendering operations is performed by the graphics device substantially in parallel with receiving the 3D scan data.

9. A machine-readable medium having executable code to cause a machine to perform a method for generating DRR images, the method comprising:

loading a volume rendering program into a graphics device, the volume rendering program representing a predetermined algorithm according to a non-linear attenuation model; and

in response to 3D scan data, invoking the graphics device to executing the volume rendering program to perform at least a portion of volume rendering operations on at least a portion of the 3D scan data, including modifying the 3D scan data according to the predetermined algorithm to compensate for a difference between a first attenuation of an object with respect to a second attenuation associating a known intensity of the object.

10. The machine-readable medium of claim 9, wherein the volume rendering operations further comprise:

casting a plurality of hypothetical rays through the modified 3D scan data from at least one of the known intensity, a known orientation, and an angle;

integrating the 3D scan data along with each hypothetical ray; and

projecting integrated values of 3D scan data onto an image plane.

11. The machine-readable medium of claim 9, wherein the 3D scan data comprises at least one of CT, MRI, PET, and ultrasound scan data.

12. The machine-readable medium of claim 9, wherein the 3D scan data comprises a plurality of CT numbers representing an image intensity of corresponding 3D CT voxels, wherein each CT voxel represents a corresponding volume element of the object, and wherein each CT number represents an attenuated intensity of an X-ray CT beam that has been generated at a CT scan energy level and that has traversed the corresponding volume element of an anatomical region.

13. The machine-readable medium of claim 12, wherein each CT voxel is disposed within one of a plurality of axial voxel slices, each axial voxel slice representing a corresponding axial slice of the object.

14. The machine-readable medium of claim 13, wherein integrating the CT numbers comprises:

performing a bi-linear interpolation for the CT voxels encountered by the ray; and

for each voxel of interest, performing a one-dimensional polynomial interpolation over the voxel of interest and for voxels on each adjacent voxel slice.

15. The machine-readable medium of claim 9, wherein the predetermined algorithm is performed according to a following algorithm:

C(x,y,z)=aC₀(x,y,z)e^{bC0(x y z)}

wherein C(x,y,z) represents the modified CT number of a 3D CT voxel having a location (x,y,z), wherein a and b represent weighting coefficients, and wherein C₀(x,y,z) represents the un-modified CT number based on a linear attenuation model of a 3D CT voxel having a location (x,y,z).

16. The machine-readable medium of claim 9, wherein the at least a portion of volume rendering operations is performed by the graphics device substantially in parallel with receiving the 3D scan data.

17. A data processing system, comprising:

a processor;

a memory coupled to the processor via a bus; and

a graphics device coupled to the bus, wherein the memory stores instructions when executed from the memory, cause the processor to

load a volume rendering program into the graphics device, the volume rendering program representing a predetermined algorithm according to a non-linear attenuation model, and

in response to 3D scan data, invoke the graphics device to executing the volume rendering program to perform at least a portion of volume rendering operations on at least a portion of the 3D scan data, including modifying the 3D scan data according to the predetermined algorithm to compensate for a difference between a first attenuation of an object with respect to a second attenuation associating a known intensity of the object.

18. An apparatus for generating DRR images, comprising:

means for loading a volume rendering program into a graphics device, the volume rendering program representing a predetermined algorithm according to a non-linear attenuation model; and

means for, in response to 3D scan data, invoking the graphics device to executing the volume rendering program to perform at least a portion of volume rendering operations on at least a portion of the 3D scan data, including modifying the 3D scan data according to the predetermined algorithm to compensate for a difference between a first attenuation of an object with respect to a second attenuation associating a known intensity of the object.

19. A computer implemented method for enhancing DRR images, the method comprising:

loading a DDR enhancement routine into a graphics device, the DDR enhancement routine representing a predetermined algorithm according to a non-linear filtering model; and

in response to a first DDR image data, invoking the graphics device to executing the DDR enhancement routine to perform at least a portion of the non-linear filtering operations on at least a portion of the first DRR image to generate a second DRR image enhanced from the first DRR image, including identifying brightest pixels in two different adjacent neighborhoods having different sizes in the first DRR image.

20. The method of claim 19, wherein if a brightest value in a smaller of the two adjacent neighborhoods is greater than a brightest value of a larger of the two adjacent neighborhoods by a predetermined threshold, the method further comprises maintaining corresponding pixels and otherwise eliminating the corresponding pixels.

21. The method of claim 19, wherein the predetermined algorithm comprises a top-hat filter.

22. The method of claim 21, wherein the top-hat filter is implemented using a weighted combination of image opening and closing with a predetermined structural element.

23. The method of claim 21, wherein the top-hat filter is determined based on the following algorithm:

f_e=f+w*[f−

γ

_B(f)]−

b*[φ

_B(f)−

f]

wherein f_e represents the enhanced image resulted from the top-hat filter to each pixel in the first DRR image, wherein f represents the first DRR image, wherein w and b represent weighting coefficients, wherein γ

_B(f) represents a structural element for an opening of the original image f, and wherein φ

_B(f) represents a structural element for the closing of the original image f.

24. A machine-readable medium having executable code to cause a machine to perform a method for enhancing DRR images, the method comprising:

loading a DDR enhancement routine into a graphics device, the DDR enhancement routine representing a predetermined algorithm according to a non-linear filtering model; and

in response to a first DDR image data, invoking the graphics device to executing the DDR enhancement routine to perform at least a portion of the non-linear filtering operations on at least a portion of the first DRR image to generate a second DRR image enhanced from the first DRR image, including identifying brightest pixels in two different adjacent neighborhoods having different sizes in the first DRR image.

25. The machine-readable medium of claim 24, wherein if a brightest value in a smaller of the two adjacent neighborhoods is greater than a brightest value of a larger of the two adjacent neighborhoods by a predetermined threshold, the method further comprises maintaining corresponding pixels and otherwise eliminating the corresponding pixels.

26. The machine-readable medium of claim 24, wherein the predetermined algorithm comprises a top-hat filter.

27. The machine-readable medium of claim 26, wherein the top-hat filter is implemented using a weighted combination of image opening and closing with a predetermined structural element.

28. The machine-readable medium of claim 26, wherein the top-hat filter is determined based on the following algorithm:

f_e=f+w*[f−

γ

_B(f)]b*[φ

_B(f)−

f]

wherein f_e represents the enhanced image resulted from the top-hat filter to each pixel in the first DRR image, wherein f represents the first DRR image, wherein w and b represent weighting coefficients, wherein γ

_B(f) represents a structural element for an opening of the original image f, and wherein φ

_B(f) represents a structural element for the closing of the original image f.

29. A data processing system, comprising:

a processor;

a memory coupled to the processor via a bus; and

a graphics device coupled to the bus, wherein the memory stores instructions when executed from the memory, cause the processor to

load a DDR enhancement routine into the graphics device, the DDR enhancement routine representing a predetermined algorithm according to a non-linear filtering model, and

in response to a first DDR image data, invoke the graphics device to executing the DDR enhancement routine to perform at least a portion of the non-linear filtering operations on at least a portion of the first DRR image to generate a second DRR image enhanced from the first DRR image, including identifying brightest pixels in two different adjacent neighborhoods having different sizes in the first DRR image.

30. An apparatus for generating DRR images, comprising:

means for loading a DDR enhancement routine into a graphics device, the DDR enhancement routine representing a predetermined algorithm according to a non-linear filtering model; and

means for, in response to a first DDR image data, invoking the graphics device to executing the DDR enhancement routine to perform at least a portion of the non-linear filtering operations on at least a portion of the first DRR image to generate a second DRR image enhanced from the first DRR image, including identifying brightest pixels in two different adjacent neighborhoods having different sizes in the first DRR image.